Surveying and Built Environment Vol 17(2), 7-28 December 2006 ISSN 1816-9554

Modelling 3D Scanned Data

to Visualise and Analyse
the Built Environment for Regeneration
Yusuf Arayici, Andy Hamilton1 and Pedro Gamito2

ABSTRACT
The renovation and refurbishment market is rapidly expanding in the construction industry. The
regeneration and transformation of cities from the industrial age (unsustainable) to the knowledge
age (sustainable) is essentially a “whole life cycle” process consisting of: planning, development,
operation, reuse, and renewal. Advanced digital mapping technologies are enablers for effective e-
planning, consultation, and communication of users’ views during the planning, design, construction,
and life cycle process of the built environment. Those technologies can be used to enhance productivity
gains by promoting a free-flow of information between departments, divisions, offices, and sites, and
between themselves, their contractors, and partners. Such is the case with the 3D laser scanner,
which enables digital documentation of buildings, sites, and physical objects for reconstruction, and
restoration. It also facilitates the creation of educational resources within the built environment, as
well as the reconstruction of the built environment. The use of a 3D scanner in combination with a 3D
printer provides the transformation of digital data from the captured CAD model back to a physical
model in an appropriate scale - reverse prototyping. The use of these technologies act as key
enablers of the creation of new approaches to the “Whole Life Cycle” process within the built and
human environment for the 21st century. This paper describes the research for building data integration
in the INTELCITIES project undertaken by a European consortium of researchers and practitioners
under the Framework 6 research programme to develop a prototype system for the e-City Platform in
order to pool the advanced knowledge and experiences of electronic government, planning systems,
and citizen participation from across Europe (www.intelcitiesproject.com). The scope includes capturing
the digital data of existing buildings using 3D laser scanning equipment and illustrating of how
digitised building data can be integrated with other types of city data, using nD modelling, to
support integrated intelligent city systems for enhancing the refurbishment process in the built environment.

KEYWORDS
3D Laser Scanner
Intelcities
Data Integration
Building Modelling
Prototyping
Object Recognition

1
School of Construction and Property Management
University of Salford, Greater Manchester, UK
y.arayici@salford.ac.uk, a.hamilton@salford.ac.uk
2
University LusÓfona de Humanidades e Tecnologias, Lisbon, Portugal
pedro.gamito@sapo.pt
SBE
7
 Modelling 3D Scanned Data to Visualise and Analyse the Built Environment for Regeneration
INTRODUCTION
Cities are transforming themselves from industrial
to post-industrial. This regeneration is seen as a
“Whole life cycle” process of construction
consisting of: Planning, Development, Operation,
and Reuse. In regards to cultural heritage,
intervention within the areas of historical and
architectural importance, where one finds
groupings of monumental and cultural heritage,
is a complex task that implies a great
responsibility, since inappropriate restoration can
result in irreversible damage (Yamada and
Takase, 2004). Therefore, it is fundamental to
have both detailed information and the tools and
technologies to allow for an ef fective
manipulation and operation of data generated
during the complete life cycle of an historical
landscape.
Virtual 3D models of monuments and structures
that have largely disappeared offer great
potential. The use of 3D digitization and
modelling in documenting heritage sites has
increased significantly over the past few years.
This is mainly due to advances in laser scanning
techniques, 3D modelling software; image-based
modelling techniques, computer power, and
virtual reality (Balletti, et al, 2004), (Arayici and
Hamilton, 2005). The difficulty in visualising 2D
plans of building elements and components
brings about errors in interpreting design
specifications in terms of accuracy, and
completeness. Besides, accessibility is another
major problem encountered by surveyors. A
number of strategies has been set out to improve
efficiency and the quality of design within the
built environment. Capturing and modelling the
3D information of the built environment is still a
big challenge. A number of techniques and
technologies are now in use. These include
Electronic Distance Measurement (EDM), Global
Positioning System (GPS) and photogrammetric
application, and remote sensing applications.
However, the 3D laser scanner comes with a
stronger advantage compared to the above
applications, as it is not limited to the surveying
SBE
8
field. A key objective is to develop existing
systems, such as Geographical Information
Systems (GIS), Virtual Reality (VR) tools, and
decision support systems for buildings, and to
employ city planning for use in urban
regeneration and integrating all with the laser
scanner to attain the benefits of working in a
holistic environment. The new integrated system
will facilitate a holistic approach to problems
and thus have a much greater functionality than
the individual sub systems.
In the next section, the 3D laser scanning
technology, its features, advantages, and
disadvantages will be explained.
3D LASER SCANNER
TECHNOLOGY
Recently, new instruments were introduced in the
field of surveying that are able to acquire portions
of land and objects of various shapes and sizes.
These instruments, based on laser technology,
are commonly known as terrestrial laser scanners.
While laser scanner instruments based on the
triangulation principle and high degrees of
precision have been widely used since the
1980s, TOF (Time of Flight) instruments have
been developed for metric survey applications
only in the last five years (Bornaz and Rinaudo,
2004, Boehler and Marbs, 2002).
These types of laser scanners can be considered
highly automated total stations. They are usually
made up of a laser, which has been optimised
for high speed surveying, and a set of mechanisms
that allows the laser beam to be directed in space
at a range that varies according to the instrument
that is being used. For each acquired point, a
distance is measured on a known direction: X, Y,
and Z coordinates of a point can be computed
for each recorded distance direction. Laser
scanners allow millions of points to be recorded
in a few minutes. Because of their practicality and
versatility, these kinds of instruments are today
widely used in the field of architectural,
 Surveying and Built Environment Vol 17(2), 7-28 December 2006 ISSN 1816-9554
archaeological, and environmental surveying
(Valanis & Tsakiri, 2004).
This innovation is significant because it has the
potential to solve problems that have always been
associated with the design and construction of
existing buildings with reuse goals. For example, it
can provide faster, better quality, and more precise
analysis and feature detection for building surveys.
Besides, in the built environment, the use of the 3D
laser scanner enables digital documentation of
buildings, sites, and physical objects for
reconstruction and restoration. It also enables the
creation of educational resources within the built
environment, including cultural heritage.
Furthermore, it can provide reverse engineering in
construction when existing buildings are
redeveloped. Producing building design and
Computer Aided Design (CAD) models and VR
models from an existing building by means of the
laser scanner will facilitate the communication
between stakeholders through 3D visualization and
facilitate an analysis of the latest conditions of the
buildings. Besides, it has also the potential to
accurately record inaccessible and potentially
Table 1: Advantages and disadvantages of 3D laser scanning technology
3D Laser Scanning
Advantages
Applicable to all 2D and 3D surfaces
Rapid 3D data collection
Very effective due to large volumes of data collected at a
predictable precision
Ideal for all 3D modelling and visualisation purposes
3D position and surface reflectance generated can be
viewed as an image
Rapidly developing survey technology
Extensive world-wide research and development
currently undertaken
hazardous areas. Consequently, it facilitates “virtual
refurbishment” of buildings and allows the existing
structure and proposed new services to be seen in
an effective manner (Ahmed, et al, 2004).
In the research undertaken for the INTELCITIES
project, laser scanner technology was
brainstormed in a SCRI workshop in October
2004, in a Construct IT workshop in November
2004, and in three interviews with two architects
from Japan and Spain and one building surveyor
from the UK. The following potential beneficiaries
is described.
As a result, this study has potential benefits and
practical applications to the construction industry.
It can provide better support for the evaluation and
visualisation of building maintenance works so that
informed policies can be effectively targeted. These
interviews can be supported by other research
activities, such as the use of 3D laser scanners to
mark the construction schedule and as-built progress
(Shih and Wang, 2004). The advantages and
disadvantages of this technology are shown in the
table below (Arayici, et al, 2004).
Disadvantages
Some systems do not work in sun or rain
Large 3D data sets require post-processing to produce a
useable output
Difficulty in extracting the edges examples from indistinct
data clouds
Output requires manipulation to achieve acceptable
recording quality
No common data exchange format currently in use
Difficult to stay up-to-date with developments
Hardware expensive and sophisticated software required
to process data
SBE
9
10
SBE
Figure 1: Potential use and benefits of laser scanner in the built environment
Deformation
Modelling 3Dand

Planning Design Construction Maintenance

Scanned

• As a communication tool • As a communication tool

Sorptivity

Building Services (Professional) (Professional)

- Electricity • As a general check tool of finish
- Air Conditioning (Professional)
DataoftoRecycled

- Health & Hygiene

• As a presentation or advertisement • As an easy measurement tool (level • As a communication tool of the spot • As a presentation tool for the
for real estate of design) (Professional) layperson
VisualiseAggregate

Interior & • As an easy modeling method to • As a modeling tool of space planning

and Analyse

Workplace explain the structure

• As a tool of explanation for concrete
Concrete

imaging (client/supplier)
the Built

• As an easy measurement tool (level • As a study modeling (Building • As an alternative of model rooms • As a deterioration measurement tool
Building of first diagnosis) Design) (level of first diagnosis)
Produced

- New Construction • As an easy modeling method to • As a communication tool to form

Environment

- Renewal explain the structure consensus

by Two

• As a tool to conserve the date of

- Conservation
withdrawal
forStage

• As an easy measurement tool (site)

Regeneration

Site
Mixing Approach

• As a tool to experience landscape • As a study modeling (Urban • As a communication tool • As a tool to conserve the data of the
City Planning Design) (Professional) memories, renewal, and history of
• As a tool to examine landscape from
- Wide Area various perspectives • As a communication tool to form city
- Region consensus
- Community • As a communication tool
(Professional)
 Surveying
Surveyingand
andBuilt
BuiltEnvironment
EnvironmentVol
Vol17(2),
17(2),15-26 December 2006
7-28 December 2006 ISSN
ISSN 1816-9554
1816-9554

INTELCITIES PROJECT RESEARCH METHODOLOGY

The INTELCITIES (Intelligent Cities) Project is a
research and development project that aims to
help achieve the EU policy goal of a knowledge
society. The INTELCITIES project brings together
the combined experience and expertise of key
players from across Europe, focusing on e-
Government, e-Planning and e-Inclusion, e-Land
Use Information Management, e-Regeneration,
Integration and Interoperability, Virtual Urban
Planning, etc. (www.intelcitiesproject.com).
This paper focuses on certain tasks related to
the laser scanner research in the e-Regeneration
work package of the project. These tasks are:
 Building data capture: Captures the digital
data of existing buildings using 3D laser
scanning equipment. Shows how this data
can be used as an information base to
enhance the refurbishment process.
 Building data integration: Build the concept
nD modelling system and relate to other data
structures, including relational databases, to
illustrate how data can be integrated to
support intelligent city systems.
In the next section, the research methodology for
how to achieve the above tasks will be explained.
Figure 2 illustrates the research methodology of
the rest of the paper, based on which is
considered the vision for the use of 3D laser
scanners and their integration with various systems
to enhance the refurbishment process.
Figure 2: The research methodology that describes the structure of the paper
This section describes the various stages
adopted for the elaboration and evaluation
process of the building data integration system
based on the research undertaken in the
INTELCITIES project. The research strategy is a
case study using a prototyping research
approach. The case studies of the project are
the Jactin House building under refurbishment
in East Manchester and the Peel Building on
the campus of the University of Salford. The
goal of the research methodology is to describe
a systematic mechanism to achieve the tasks
as cited in the previous section. The evaluation
process was carried out in four stages, as
illustrated in Figure 2. According to the figure,
at the first stage, the vision for a conceptual
system of building data integration is described.
The concept covers the integration of laser
scanner technology with other technologies and
systems, such as virtual reality (VR), GIS, CAD,
the nD modelling database, and so on. At the
second stage, the integration of the laser
scanner technology with CAD systems, VR
toolkits, and 3D printers is further elaborated
on for modelling real world data, such as
buildings, by means of a prototyping technique.
At the third stage, the concept of a visual
decision making support system through the
integration of a laser scanner with GIS is
elaborated on. Last, the research findings and
progress on integration with nD modelling
database are presented, which explains the nD
modelling database, pattern matching, and

1. Describe the Vision for the Conceptual System of Building Data Integration

2. Identify the prototyping approach for real world data modelling for integration with 3D printers and VR
equipment according to the vision

3. Describe the system architecture of a Visual Decision Making Support System for the integration of laser
scanners with GIS

4. Identify the Strategy for the integration with an nD modelling database that is described in the vision

SBE
11
 Deformation
Modelling 3Dand Sorptivity
Scanned DataoftoRecycled
VisualiseAggregate Concrete
and Analyse
the Built
object recognition approaches for data
modelling from point cloud data to object-
oriented information. This section describes the
approaches adopted for this process.
The Vision of Conceptual System of
Building Data Integration
The scope for the conceptual system of building data
integration in the INTELCITIES project is depicted in
Figure 3 below. It includes the integration of laser
scanner technology with various systems, such as CAD
systems for 2D and 3D CAD plans, and the use of
the virtual environment tools, such as workbench, a
VR projection system, a video conferencing system
for visualisation and communications, along with a
3D printing system for prototyping, a GIS for a visual
decision making support system, and the nD modelling
repository for storing the Industry Foundation Classes
(IFC) building information. These are produced with
a laser scanning system in a database that embraces
information in various formats for future use during the
Figure 3: Use case diagram for the conceptual modelling
3D Laser Scanning
Implementer
Developing digital building
information
Developer
Integration of the Information
<<include>>
<<include>>
Visualisation
Public
SBE
12
Produced
by Two
Environment forStage Mixing Approach
Regeneration
refurbishment process.
The use case diagram in Figure 3 illustrates the
building data integration system. The use cases
were used for the conceptual modelling of what
the system should do from the user’s point of view.
It also indicates a number of actors who will be
engaged in system. While the implementer
captures the building data and develops VR
models of the buildings, the developer will
integrate the VR models of the laser scanned
data with various systems to make it ready for a
variety of users for building refurbishment and
planning. For example, members of the public
can visualise the model, and the building
surveyor can put the CAD models of the building
up for refurbishment, while the planning officer
and heritage manager use 3DGIS (3
Dimensional GIS) for context and environmental
analysis and specialist produce physical models
from digital models and IFC data for the nD (n
Dimensional) modelling database.
Multiple Scan alignment
<<include>>
<<include>>
<<include>> Generate Polygonal Model
Building Surveyor
Feature extraction for CAD
<<include>> implementation
<<include>>
integration with the nD Modelling
<<include>> Database Specialist
integration with 3D printer for the
Physical building Mod...
Planning Officer
Integration with 3D GIS&VR
system for decision making
Heritage Manager
 Surveying
Surveyingand
andBuilt
BuiltEnvironment
According to the use case model above, when
the concept of a building data integration system
is implemented, a variety of actors can benefit
from the system. For instance, in regards to cultural
heritage, many historical sites are slowly
deteriorating due to exposure to the elements.
Although remedial efforts can reduce the rate of
destruction, a digital model of the site will preserve
it indefinitely. Models of historical artefacts also
allow scientists to study the objects in new ways.
For example, archaeologists can measure artefacts
in non-contact fashion, which is useful if an object
is fragile. Also, digital models allow objects to
be studied remotely, saving time and travel
expenses and enabling more archaeologists to
study them. Besides, the heritage manager and
planning officer can utilise the system by using
the building information captured by means of a
laser scanner in the decision making process and
verifying the coherence between cultural heritage
and city planning interventions. Often the
materials, constructive pathologies, and systems
are insufficient or deficient for archiving and the
sharing of information in the documentation of
historic buildings, sites, and objects for
refurbishment.
In regards to deformation and inspection,
bridges, for example, need to be surveyed
periodically to determine whether significant
settling or other movement has occurred. A 3D
model of the site allows engineers to make the
same measurements with the equivalent accuracy
in a fraction of the time.
With regards to reverse and rapid prototyping,
modelling from reality can be used to reverse
engineer real objects, for which a CAD model
may not be available or has never existed. The
process of modelling from reality imports the real
world objects into the computer environment,
creating a digital model that can be edited with
a CAD program and then creating a physical
model of the same object. This will be useful for
the communication between the stakeholders in
the refurbishment process, in particular the client
and architect, and for the publicity of the real
EnvironmentVol
Vol17(2),
17(2),15-26 December 2006
7-28 December 2006 ISSN
ISSN 1816-9554
1816-9554
object or building under refurbishment.
Regarding architecture and construction,
architects frequently need to determine the “as
built” plans of building or other structures, such
as industrial plants, bridges, and tunnels. A 3D
model of a site can be used to verify that it was
constructed according to specifications. When
preparing for a renovation or a plant upgrade,
the original architectural plans might be
inaccurate, if they exist at all. A 3D model allows
an architect to plan a renovation and test out
various construction options.
In regards to engineering, the process of modelling
from reality offers an efficient alternative for
engineers and surveyors. For example, engineers
can use the information that is converted into IFC
in the building life cycle for regeneration.
The following section describes the integration
of laser scanner technology into 3D printer and
Virtual Environment tools.
Real world data modelling
The main purpose of the conceptual model
described above is to help one think about real
world problems. The same dimensions of
integration in the conceptual model can be
implemented by using the BUHU (Research Institute
of Built and Human Environment) equipment.
The spatial data can be obtained by using a
3D scanner. By post-processing the captured
spatial data, outputs for different purposes can
be obtained, such as CAD modeling, physical
modeling by prototyping, and visualization on
different platforms. It is depicted in Figure 4.
The use of these technologies is a key enabler in
the creation of an integrated system to capture,
process, and display 3D information.
For example, these technologies can be used
for civil engineering and environmental analysis.
It permits the user to acquire irregular point clouds
of land areas, rivers, and infrastructure in a fast
SBE
13
 Deformation
Modelling 3Dand Sorptivity
Scanned DataoftoRecycled
VisualiseAggregate Concrete
and Analyse
the Built
Figure 4: Integration of spatial data with 3D printer and the virtual environments
and cheap way. The use of raw laser scanner
data requires orientation and filtering procedures
to generate a 3D model of the surveyed object.
It is possible to automatically derive a set of
geometric information from this 3D model that is
useful for a variety of particular engineering and
environmental applications, such as dense Digital
Terrain Model (DTM) generation, sections and
profiles, contour maps, volumes, and so on
(Bornaz, et al, 2002).
Integration of spatial data can also provide faster,
better quality, and more precise analysis and
feature detection for building surveys (Arayici,
et al, 2004). One of the major problems that
building sur veyors have encountered is
inaccessibility. That is to say, on some occasions,
they have to measure the details of building and
relevant architectural details on the walls for
example, particularly when developing elevation
plans. In these circumstances, building surveyors
are required to estimate the measurements of
these details, which is very much likely to involve
SBE
14
Produced
by Two
Environment forStage Mixing Approach
Regeneration
a variety of errors regarding the accuracy and
the actual design of a building due to a wrong
assumption or judgement.
The use of a 3D laser scanner has the potential
to benefit building surveyors and their clients in
terms of accuracy, speed, and productivity in plan
preparation, and then to extend the range of
services offered through modelling applications.
The output of feasibility studies would improve
immeasurably through modelling in areas, such
as those concerning disabled access and fire
safety. With current demand for whole life
costing, asset management planning, and
database application, building surveyors are
called upon to link information in CAD files to
database files. The possibility of linking this
information to 3D models would suggest all sorts
of additional benefits. For example, producing
a computer model of historical building with 3D
laser scanner technology can only take a week,
and this includes scanning the interior and exterior
of a building in two days and post-processing
 Surveying
Surveyingand
andBuilt
BuiltEnvironment
the point cloud data in three days. For this work,
only two people would be required to both
complete the initial scans and post-process the
raw data. Consequently, this technology is
significantly faster and more effective than any
other surveying application in use. However, the
post-processing part may continue depending
on the end product sought. For example, if it is
aimed at rapid prototyping, another two weeks
of work will be necessary to complete it.
Rapid Prototyping
The term rapid prototyping refers to a class of
technologies that automatically construct physical
models from Computer Aided Design (CAD)
data. These “three dimensional printers” allow
designers to quickly create tangible prototypes
of their designs, rather than just two dimensional
pictures. Such models have numerous uses. They
make excellent visual aids for communicating
ideas with co-workers or clients. In addition,
prototypes can be used for design testing. There
are research activities in rapid prototyping for
the reverse engineering and design of buildings
in particular historic buildings. For example, Alves
and Bartolo (2006) proposed integrated
computational tools for virtual and physical
automatic construction using the human vision
approach and 2D photos for 3D computer
modelling. However, in this paper rapid
prototyping is conducted in accordance with the
ground-based 3D laser scanner data.
Although several rapid prototyping techniques
exist, all employ the same basic five-step process.
These steps are as follows:
1) Create a CAD model of the design: First, the
object to be built is modelled using a
Computer Aided Design (CAD) software
package, such as Microstation, which is used
together with Polywork point cloud modeller
in order to create a CAD model from the laser
scanned data of real objects, such as buildings.
First, a model is lined up against the reference
for the coordinate system of Polywork software
EnvironmentVol
Vol17(2),
17(2),15-26 December 2006
7-28 December 2006 ISSN
ISSN 1816-9554
1816-9554
and co-planar surfaces with different distances
and depths are defined according to the
model’s facade. Some planes are parallels
to the model and others perpendicular. Next,
vertices on the corresponding surfaces are
selected and projected to the co-planar
surfaces previously defined. With this process
of projection of points onto the co-planar
surfaces, the edges are also defined. Then,
the mesh model is optimised to rectify the
imperfection in the triangulation that existed
in the initial model. Cross-sections are inserted
into the co-planar surfaces to obtain the feature
lines of the model. Last, the cross-sections are
exported in dxf and then imported into
Microstation. The result in Microstation is a
regular 3D CAD model of the building.
2) Convert the CAD model to STL format:
To e s t a b l i s h c o n s i s t e n c y, t h e S T L
(Stereolithography) format has been adopted
as the standard of the rapid prototyping
industry. This format represents a three-
dimensional surface as an assembly of planar
triangles. The STL file contains the coordinates
of the vertices and the direction of the outward
normal of each triangle. Because STL files
use planar elements, they cannot represent
curved surfaces precisely. Increasing the
number of triangles improves the
approximation, but at the cost of bigger file
size. Large, complicated files require more
time to pre-process and build, so a designer
must balance accuracy with manageability
to produce a useful STL file. Since the .stl
format is universal, this process is identical to
all of the rapid prototyping techniques.
3) Slice the STL file into thin cross-sectional layers:
In the third step, a pre-processing program
prepares the STL file to be built. Several
programs are available, and most allow the
user to adjust the size, location, and
orientation of the model. The pre-processing
software slices the STL model into a number
of layers from 0.001 mm to 0.07 mm thick,
depending on the build technique. The
SBE
15
 Deformation
Modelling 3Dand Sorptivity
Scanned DataoftoRecycled
VisualiseAggregate Concrete
and Analyse Produced
the Built by Two
Environment forStage Mixing Approach
Regeneration

program may also generate an auxiliary
structure to support the model during the
construction. Supports are useful for delicate
features, such as overhangs, internal
cavities, and thin-walled sections.
4) Construct the model one layer on top another:
The fourth step is the actual construction
of the part. Using one of several
techniques, rapid prototyping machines
build one layer at a time from polymers or
powdered metal.
5) Clean and finish the model: The final step is post-
processing. This involves removing the prototype
from the machine and detaching any supports.
Some photosensitive materials need to be fully
cured before use. Prototypes may also require
minor cleaning and surface treatment.
Figures 5 to 10 show the Jactin House example,
which is the case study building in East
Manchester, for VR models produced from the
building data captured by the 3D Laser scanner
and CAD models extracted from the VR models.

Figure 5: 12 scans were conducted around Jactin house; the picture below is one of the scans

Figure 6: The individual scans are processed to create the VR model below

SBE
16
 Surveying and Built Environment Vol 17(2), 7-28 December 2006 ISSN 1816-9554

Figure 7: Using co-planar surfaces and creating sharp edges from curves in the model in Figure
6, the CAD model below is developed using Polyworks modeller and Microstation
software

Figure 8: Polygonal models for printing can be generated from the CAD model above

Figure 9: The images below are samples of the Jactin House model

SBE
17
 Modelling 3D Scanned Data to Visualise and Analyse the Built Environment for Regeneration
Figure 10: Peel building data captured by the laser scanner is textured and presented
The sequence of figures from Figure 5 to Figure
10 also illustrates the process of data from real
world to digital modelling and to physical
modelling, which is one of the dimensions of
integration between the laser scanner and the
3D printer depicted in the conceptual model of
building data integration system in Figure 3.
As illustrated in Figure 4, the system can collect
data from the real world and manipulate it for
integration. For example; the geometric information
can be obtained using a 3D scanner. By editing
and integrating the data, a model of the object
(building/urban area) could be created to be used
in the decision-making process. The model can be
presented in different ways, for example, as a
physical model or VR projection system for reverse
and forward prototyping, which is shown in Figures
5, 6, 7, 8, and 9. An example of visualization in
the VR projection system is illustrated in Figure 10.
This implementation depicted in Figure 10 is the
dimension of integration of the laser scanner with
the VR environment tools shown in the concept
model of building data integration system in
Figures 3 and 4.
SBE
18
In the next section, another dimension of the
integration of a laser scanner with GIS will be
discussed for a Visual Decision Making Support
System in regeneration.
An Approach for a Visual Decision
Making Support System
The aim of this section is to describe a framework
that includes a geospatial database, such as the
nD modelling database, the 3D laser scanner,
and a series of analytical tools that will enable
various stakeholders in the regeneration process
to make decisions relating to building maintenance
works (Ahmed, et al, 2004). This framework is
depicted in Figure 12, and includes:
1) The development and population of a geo-
spatial project database with the digital data
of existing buildings captured with laser
scanning equipment.
2) The analysis of complex building information
maintenance options within a knowledge
repository environment, in which digital
building data that was captured by the laser
 Surveying and Built Environment Vol 17(2), 7-28 December 2006 ISSN 1816-9554
scanner is retrieved with a 3D GIS system for
the analysis.
3) The visualisation of project information through
a range of different interconnected graphic
windows. The laser scanner VR model can
be visualised on different platforms.
The geospatial database describes the
geometries of both the building frame and its
components. Simple open geometric descriptions
are used, but each entry will also be associated
with the data on inventory information, such as
name, supplier, date installed, replaced, number
of replacements, etc (Ahmed, et al, 2004).
The system will enable the capture of geo-spatial
data using a laser scanner. This data will be
processed and building frame and components
will be determined in a CAD environment.
Inventory information relating to each frame and
component will also be captured within the
relational structure of the database. Such
information will be accessible in real-time with
Figure 11: Structure of a visual decision making support system
some of the attributes (e.g. component supplier
information) made available through the
communication layer.
GIS software will be used to generate and
analyse thematic developments relating to the
building properties and associated refurbishment
management strategies. Such software will
retrieve the building information from the
database through the communications layer,
which provides users situated in various locations
with access to the database.
The VR environment will be created in a Virtual
Reality Modelling Language (VRML) interface.
The systems are developed based on the sharing
of construction information via a central object
orientated project database. Typically, the CAD
application allows a user to create and
manipulate the components of a facility. The
components are stored as instances of classes
in an object-oriented database. These instances
are read by the VRML interface in order to create
a 3D view of the facility, which gives the user a
SBE
19
 Modelling 3D Scanned Data to Visualise and Analyse the Built Environment for Regeneration

better environment for navigation and
walkthrough. Therefore, the results of the spatial
analysis obtained within the 3-D GIS environment
can be evaluated in real-time with the options of
viewing building refurbishment alternatives
(Mahdjoubi and Ahmed, 2004).
In the next section, the research methodology
for the actual development and implementation
of the conceptual system of building data
integration will be described.
Integration with the nD modelling
Repository
The building data captured by the laser scanner
is processed to generate information that will be
stored in the nD modelling database in IFCs
(Industry Foundation Classes). The information
stored in the database will be available for the
decision making process and be reused in the
refurbishment process in the future. To make
progress in modelling, conversion issues of 3D
scanned data need to be solved. There are certain
barriers related to laser scanner data. For example,
there are no standard formats for the distribution
of scanned data. This leads to issues relating to
compatibility, the exchange of information, and
data archiving. These barriers can be overcome
through data conversion from scanned data to
standard data formats such as IFC (IFC2x2,
2005). Once this conversion is achieved, it is
time to store this information characterised by IFC
schema in the nD modelling database.
Figure 12 shows the n-dimensional conceptual
urban data model. Entities such as ‘road, railway,
pollution, and IFC Building’ have been
described. The core of the conceptual data
model is the abstract entity of ‘administrative
boundary’. An administrative boundary is
defined as the limit of responsibility area. A
‘country’ could be the largest administrative area.
Countries are composed of counties, cities,
districts, and parcels. A parcel can also be
divided into smaller units called ‘partition
parcels’. The attributes of the administrative
boundary are all inherited by these sub classes.
The ‘administrative boundary’ classes, such as
countries, counties, cities, districts, and parcels,
have relations with geographical entities, such
as roads, railways, water elements, heritage
elements, etc. A parcel may have a building on
it. IFCs were used to model the building related
data. An administrative boundary may also
attract crime and pollution. A city, a parcel, or
a user-defined boundary unit may be selected
for the air, noise pollution simulations (Tanyer, et
al, 2005a, 2005b).

In this section, the integration concept with the

nD modelling database will be elaborated on.
The nD modelling database is a multi-dimensional
data storey to keep information for various
systems in the building life cycle. The aim of the
database development is to design, build, and
evaluate the components of an interactive virtual
urban planning environment as part of the e-
City platform. The nD modelling database is the
base of this interactive virtual urban planning
environment (Hamilton, et al, 2005).

SBE
20
 Surveying and Built Environment Vol 17(2), 7-28 December 2006 ISSN 1816-9554

Figure 12: Conceptual urban data model (Source: Tanyer, et al, 2005)

SBE
21
 Modelling 3D Scanned Data to Visualise and Analyse the Built Environment for Regeneration

From 3D Scanned Data to nD Modelled Data
The research to date about object recognition
from the laser scanned data has focused on a
pattern matching approach. Pattern Matching
addresses issues of searching and matching
strings and more complicated patterns, such as
trees, regular expressions, graphs, point sets, and
arrays (Johnson, 2002). The diagram below
shows the pattern matching approach in the
INTELCITIES project for object recognition from
the laser scanned data.
3D CAD models using semi-automated
techniques were extracted from the polygonal
mesh model developed in the laser scanner
system. The extracted CAD model is presentable
in any commercial CAD software. According to
Figure 13, the pattern matching interface being
designed will invoke the 3D CAD model to its
display screen. Highlighting any building frame
in the model will enable the pattern matcher to
define geometric features as criteria for matching
processes such as shape, sides, width, height,
thickness, line type, line thickness, line colour,
and so on. Besides the automatic feature
recognition, users can also input further criteria
into the matching process. However, this process
is being completed manually in Microstation
Triforma for now because the pattern matching
algorithm definition is still under investigation.
The IFC building model of Jactin House is
presented in Figure 14.
Microstation triforma employs a single building
model concept. All information about a building
is recorded in a 3D model. Traditionally, a given
door in a building would be drawn in at least
three or four places (plan, building elevation,
building section, interior elevation, etc). In the
single building model, it is constructed once and
these various drawings are later extracted
automatically. The single building model requires
building objects, which are defined, edited, and
stored in the triforma library.

Figure 13: Data Conversion approach for laser scanned building data

Laser scanner Produces 3D CAD 3D CAD

system Model

Invoke the

Search, match, and retrieve

object attributes in the library
Triforma Pattern matching
Object Library interface

Assign the object attributes to the 3D

CAD Model

Save and store

nD Modelling the model in the IFC data model Object-Oriented
database CAD Model

SBE
22
 Surveying and Built Environment Vol 17(2), 7-28 December 2006 ISSN 1816-9554
The pattern matcher will access the triforma
library with the criteria to do “search and match”.
Two different types of matching can be done: 1)
the exact pattern matching, and 2) the
approximate pattern matching. Exact pattern
matching consists of finding the exact pattern. In
the case of approximate pattern matching, it is
a generalisation of the pattern sought and a
determined number of differences between the
pattern sought and the objects found in the library
is allowed.
There will be three types of pattern form: simple,
constrained, and variable. All three forms can
be accommodated by the pattern matcher
interface. However, the pattern matcher will
handle them differently to improve efficiency.
Simple patterns are simple matches on values.
The pattern describes which values must be
included in the search. For example, the pattern
matcher tries to match an instance of a square
(a frame) and it finds square-1, square-2, square-
3, and square-4, which are all the squares in
the example database.
In case of a constrained pattern form, it is possible
to specify a constraint on a value instead of a
specific value. To do so, a function to be
provided will test the constraint. For example,
Figure 14: The building details are defined as triforma objects
the pattern matcher matches the frame to triangle-
1, triangle-2, triangle-3, and triangle-4. The
constraint (value <4) limits the value of sides to
less than 4.
In respect to variable types of pattern form,
patterns may also include variable references.
Variables in patterns can be used to relate values.
The first occurrence of a variable in a pattern
causes the value in a frame to be bound to that
variable. The second occurrence forces a
constraint that the value will be equal to the value
bound to the variable. For example, pattern
(width a, height a) describes objects with widths
equal to their heights.
Variables may also be used in constraints to relate
values through more complex relations than
simple equality. For example, pattern (width a,
height {value >a}) describes objects with heights
greater than width. Variables may also be used
to relate values across patterns. Since each
pattern refers to values in a single frame, relating
values across frames requires multiple patterns
(Yates, 2004). The function, multiple-pattern
match, searches for frames that match a set of
related patterns. The multiple-pattern match returns
a list of sub-lists in which the sub-lists contain the
value of the variables followed by the frames
SBE
23
 Modelling 3D Scanned Data to Visualise and Analyse the Built Environment for Regeneration
that match each pattern within the set. The
variables can also be related across frames with
constraints. For example, pattern (width a,
height {value >a}, sides 4) describes all objects
with heights greater than the width and with 4
sides.
As a result of various matching processes, object
recognition can be worked out for the interested
building frames in the 3D CAD model. Attributes
of the objects matched in the library will be assigned
to the building frame in the CAD model, which
result in the building frames to be defined as building
objects. Subsequently, an object-oriented (OO)
CAD model will be obtained. This OO CAD model
will be mapped into IFC schema to be saved and
stored in the nD modelling database.
The Object Recognition Approach
To progress in the area of modelling, conversion
issues of 3D scanned data need to be solved.
The general problems of 3D data integration
in the built environment were considered by
Wang and Hamilton (2005). This section
specifically addresses the problem of object
recognition. This is important in realizing the
visualization of 3D models of the built
environment because visualization will be
undertaken with the real characteristics and
features of real world objects, as opposed to
raw 3D models of laser scanned data, which
is difficult to handle for many visualization
software applications due to large file sizes.
As a result, data modelling and analysis of laser
scanned data can also be done more effectively
and precisely. This also facilitates the use of
CAD and GIS software, which will not normally
support raw scanned 3D models.
There are certain barriers. For example, there
are no standard formats for the distribution of
scanned data. This leads to issues relating to
compatibility, the exchange of information, and
data archiving. These barriers can be overcome
through data conversion from scanned data to
a standard data format, such as IFC (IAI, 2004)
or IFG (Industry Foundation Classes for GIS).
SBE
24
IFC is an international data standard model for
information sharing and interoperability in the
construction industry. On the other hand, IFG
also makes it possible to communicate relevant
intelligent information from various GIS
standards to CAD systems using IFC (Wang
and Hamilton, 2005).
In the INTELCITIES research programme, we
have an approach for object recognition through
the simultaneous recognition of multiple objects
in the scenes of polygonal mesh models
produced from the laser scanner point cloud
data, which contains clutter and occlusion.
Recognition is based on matching surfaces by
matching points using spin image representation.
Spin imaging is a data level descriptor that is
used to match surfaces represented as a surface
mesh (Johnson and Hebert, 1999). Figure 15
below describing the process of object
recognition from the laser scan data was
developed based on a spin image definition from
Johnson and Hebert (1998, 1999).
Through surface matching, an object can be
identified in a scene of laser scanned data by
comparing a targeted surface to an object
surface stored in a database. When the object
surface is matched, an interrelation can be
established between something known (the
object) and something unknown (the scene of
laser scanned data). As a result, information
about the world is acquired.
Surface matching is based on matching
individual surface points in order to match
complete surfaces (Johnson and Hebert, 1999).
By matching points, the problem of surface
matching is broken down into many smaller
localized problems of point matching. As a
result, matching points provides a method for
handling clutter and occlusion in surface
matching without first segmenting the scene.
Clutter points on one surface will not have
matching points on the other, and occluded
points on one surface will not be sought on the
other (Ullman, et al, 2002).
 Surveying and Built Environment Vol 17(2), 7-28 December 2006 ISSN 1816-9554

Figure 15: Object recognition from the laser scanner system

VR model of Produce scene API (does object Produce scene Object

a building spin image recognition spin image database

Match the scene spin image

with the object spin images

Assign the objects from object

spin images to the corresponding
elements in the scene spin image

Write the recognised Write the recognised

object into IFC schema objects into IFG schema

Store the data in Store the data in

the database the database

nD modelling database

To differentiate among points, 2D images
associated with each point were constructed.
Oriented points, which are 3D points with
associated directions, were used to create spin
images. An oriented point was defined at a
surface mesh vertex using the 3D position of the
vertex and surface normal at the vertex. The
surface normal at a vertex was computed by
fitting a plane to the points connected to the
vertex by the edges at the surface mesh (Johnson
and Hebert, 1998, 1999).
According to Figure 15, the polygonal model
mesh was invoked by the API (Application
Programming Interface) to be developed in
order to create a scene spin image based on a
selected scene oriented point. At the same time,
the system will access the object database to
create spin images for each object model in
the database, and these spin images are stored
in a spin image stack. Point correspondences
were then established between the selected
point and the points with the best matching spin
images on the other surface. This procedure
was repeated for many points until there were
a sizeable number of set point correspondences.
The point correspondences were then grouped
and the outliers eliminated using geometric
consistency.
This surface matching can be extended to object
recognition as follows. Each object in the object
database is represented by a polygonal mesh.
Before recognition, the spin images for all
vertices on all models were created and stored.
At recognition time, a scene point was selected
and its spin image was generated. Next, its
spin image was correlated with all the spin
images from all the objects. The best matching
model spin image will indicate both the best
matching model and model vertex after
matching many scene spin images to model

SBE
25
 Modelling 3D Scanned Data to Visualise and Analyse the Built Environment for Regeneration
spin images. The result was a simultaneous
recognition and localization of the objects that
existed in the scene (Stein and Hebert, 2005).
The following step is to assign and populate
objects to the corresponding elements in the
scene of the laser scanner VR mesh model. In
the end, a new VR model is an object-populated
mesh model. The following step is to write these
objects at the scene into the IFC schema at
building scale or IFG schema at city scale (Wang
and Hamilton, 2005). IFC-enabled CAD
software assists professionals when they plan a
building or an area with buildings by supplying
an attribute information type of window, cost
and performance, directly connected to the 3D
model. For larger scale integrated data, it is
also necessary to extract parts of the site and
building information (in IFC form) and make it
available in GIS standards.
The information characterised by either the IFC
schema or IFG schema will be stored in the nD
modelling (Lee at al, 2003) database. The nD
modelling database is a multi-dimensional data
storey for keeping information for various systems
at the building scale and urban scale. This is an
ongoing progress, and the focus of future work
will be on the algorithm definition of object
recognition for data conversion from scanned
data to nD modelled data.
CONCLUSION
This paper explained the research on the use of
laser scanner technology for the built
environment. The focus was on e-regeneration
to facilitate the building refurbishment process.
Therefore, building data captured by laser
scanners and building data integration with
various systems were addressed.
In the paper, the INTELCITIES research project
was introduced. Their scope for the use of the
laser scanner technology for the built environment
at the building scale and urban scale was
SBE
26
described. The research undertaken in the project
for 3D laser scanner technology mainly focused
on data capture, modelling, and integration.
Case studies of Jactin House and the Peel
Building were given as examples for CAD and
VR modelling and integration with other
technologies, such as the 3D printer and VR
projection system. Furthermore, the potential
benefits of using a 3D laser scanner were
addressed. Last, an approach to object
recognition for object-oriented data modelling
from the 3D scanned data was discussed. The
benefits of this approach were noted. For
example, providing information in standard
formats, such as IFC and IFG on the built
environment using the laser scanner through
object recognition approach, will enable one
to communicate with other city and building
systems, including CAD and GIS systems.
Generally speaking, the present use of laser
scanners in the built environment was criticised
as being too laborious and time consuming.
However, the approaches explained in this
paper can lead to answers for overcoming these
issues, and it is envisaged that when the data
conversion problems are solved according to
the methods explained in this paper, the use of
laser scanner VR models can be developed
quickly, which will make it feasible for intelligent
visualizations to be produced for a wide range
of applications.
REFERENCES
Ahmed V, Arayici Y, Hamilton A and Aouad G
(2004), Vir tual Building Maintenance:
Enhancing Building Maintenance Using 3D-GIS
and 3D Laser Scanner (VR) Technology,
Proceeding of European Conference on Product
and Process Modelling in the Building and
Construction Industry (ECPPM), 8-10 September
2004, Istanbul, Turkey, 29-34.
Alves NM and Bartolo PJ (2006), Integrated
computational tools for virtual and physical
 Surveying and Built Environment Vol 17(2), 7-28 December 2006 ISSN 1816-9554
automatic construction, Journal of Automation in
Construction, 15:3, 257-271.
Arayici Y and Hamilton A (2005), Modelling
3D Scanned Data to Visualise the Built
Environment, Proceedings of 9th International
Conference on Information Visualisation, IEEE
Computer Society, 6-8 July 2005, London,
509-514.
Arayici Y, Hamilton A, Gamito P and Albergaria
G (2004), The Scope in the INTELCITIES Project
for the Use of the 3D Laser Scanner, Proceeding
of ECT2004: The Four th International
Conference on Engineering Computational
Technology , 7-9 September 2004, Lisbon,
Portugal.
Balletti C, Guerra F, Vernier P, Studnicka N,
Riegl J and Orlandini S (2004), Practical
Comparative Evaluation of An Integrated Hybrid
Sensor Based On Photogrammetry and Laser
Scanning for Architectural Representation, XXth
ISPRS Congress, 12-23 July 2004 Istanbul,
Turkey, 536-541.
Boehler W and Marbs A (2002), 3D Scanning
Instruments, Proceeding of the CIPA WG 6
International Workshop on Scanning For Cultural
Heritage Recording, 1-2 September 2002,
Corfu, Greece.
Bornaz L, Lingua A and Rinaudo F (2002),
Engineering and environmental applications of
laser scanner techniques, International Archives
of Photogrammetry and Remote Sensing. ISSN:
1682-1777. Volume XXXIV part 3A+BCD-ROM.
Bornaz L and Rinaudo F (2004), Terrestrial Laser
Scanning Data Processing , XXth ISPRS
Congress, 12-23 July 2004, Istanbul, Turkey,
514-520.
Hamilton A, Wang H, Tanyer AM, Arayici Y,
Zhang X and Song Y (2005), Urban information
model for city planning, ITcon, 10:6, 55-67.
IAI ( 2002) , Inter national Alliance for
Interoperability, Available online at: http://
www.iaiinternational.org, (accessed 20 June
2005).
IFC2X2 (2005), IFC2x Edition 2 Addendum 1,
Available online at: http://www.iaiinternational.
org/iai_international/Technical_Documents/
R2x2_add1/index.html, (accessed 20 June
2005).
Johnson AE, and Hebert M (1998), Surface
Matching for Object Recognition in Complex
Three-Dimensional Scenes, Image and Vision
Computing, 16, 635-651.
Johnson AE and Hebert M (1999), Using Spin
Images for Efficient Object Recognition in
Cluttered 3D Scenes, IEEE Transaction on Pattern
Analysis and Machine Intelligence, 21:5, 433-
449.
Johnson M (2002), A simple pattern matching
algorithm for recovering empty notes and their
antecedents, Proceeding of 40th Annual Meeting
of the Association for Computational Linguistics
(ACL), July 2002, Philadelphia, 136-143.
Lee A, Marshall-Ponting AJ, Aouad G, Wu S,
Koh WWI, Fu C, Cooper R, Betts M, Kagioglou
M and Fisher M (2003), Developing a vision of
nD-enabled construction, Construct IT, University
of Salford, UK.
Lichti D (2004), A Resolution Measure For Terrestrial
Laser Scanners, XXth ISPRS Congress, 12-23 July
2004 Istanbul, Turkey.
Mahdjoubi L, and Ahmed V (2004), Virtual
Building Maintenance: Enhancing Building
Maintenance using 3D GIS and Virtual Reality
(VR) Technology’ , Conference of Designing,
Managing, and Supporting Construction Projects
through Innovation and IT solutions
(INCITE2004), February 2004, Langkawi,
Malaysia.
SBE
27
 Modelling 3D Scanned Data to Visualise and Analyse the Built Environment for Regeneration

Stein A and Hebert M (2005), Incorporating

Background invariance into Feature-Based
Object Recognition, 7th IEEE Workshop on
Application of Computer Vision, January 2005.

Shih NJ and Wang PH (2004), Point-Cloud-

Based Comparison between Construction
Schedule and As-Built Progress: Long-Range
Three-Dimensional Laser Scanner's Approach,
Journal of Architectural Engineering, 10:3, 98-
102.

Tanyer AM, Tah JHM and Aouad G (2005a),

An Integrated Database to Support Collaborative
Urban Planning: The N-Dimensional Modelling
Approach, ASCE Conference on Computing in
Civil Engineering, 12-15 July 2005, Cancun,
Mexico.

Tanyer AM, Tah JHM and Aouad G (2005b),

Towards n-Dimensional Modelling in Urban
Planning, Innovation in Architecture,
Engineering & Construction (AEC), 15-17 June
2005, Rotterdam, 503-512.

Ullman S, Vidal-Naguet M and Sali E (2002),

Visual features of intermediate complexity and
their use in classification, Nature Neuroscience,
5:7, 682-687.

Wang H and Hamilton A (2005), Data

integration issues within nD Information Modelling
for Urban Planning, 5th International
Postgraduate Research Conference, Salford, UK,
194-203.

Yamada O and Takase Y (2004), A Case Study

For the Practical Use of 3D Digital Archive of
Cultural Properties, XXth ISPRS Congress, July
2004, Istanbul, Turkey, 12-23.

Yates BR (2004), A Fast Set Intersection Algorithm

For Sor ted Sequences, Proceeding of
Combinatorial Pattern Matching conference
(CPM), July 2004, Istanbul.

SBE
28